Timing of Parathyroidectomy Does Not Influence Renal Function After Kidney Transplantation

Parathyroidectomy (PTx) is the treatment of choice for end-stage renal disease (ESRD) patients with therapy-resistant hyperparathyroidism (HPT). The optimal timing of PTx for ESRD-related HPT—before or after kidney transplantation (KTx)—is subject of debate. Patients with ESRD-related HPT who underwent both PTx and KTx between 1994 and 2015 were included in a multicenter retrospective study in four university hospitals. Two groups were formed according to treatment sequence: PTx before KTx (PTxKTx) and PTx after KTx (KTxPTx). Primary endpoint was renal function (eGFR, CKD-EPI) between both groups at several time points post-transplantation. Correlation between the timing of PTx and KTx and the course of eGFR was assessed using generalized estimating equations (GEE). The PTxKTx group consisted of 102 (55.1%) and the KTxPTx group of 83 (44.9%) patients. Recipient age, donor type, PTx type, and pre-KTx PTH levels were significantly different between groups. At 5 years after transplantation, eGFR was similar in the PTxKTx group (eGFR 44.5 ± 4.0 ml/min/1.73 m2) and KTxPTx group (40.0 ± 6.4 ml/min/1.73 m2, p = 0.43). The unadjusted GEE model showed that timing of PTx was not correlated with graft function over time (mean difference −1.0 ml/min/1.73 m2, 95% confidence interval −8.4 to 6.4, p = 0.79). Adjustment for potential confounders including recipient age and sex, various donor characteristics, PTx type, and PTH levels did not materially influence the results. In this multicenter cohort study, timing of PTx before or after KTx does not independently impact graft function over time.


Introduction
Hyperparathyroidism (HPT) is a common metabolic complication in end-stage renal disease (ESRD) [1]. Both before and after kidney transplantation (KTx), HPT has been associated with adverse patient outcomes [2,3]. Parathyroidectomy (PTx) is the treatment of choice for patients with HPT refractory to pharmacological treatment [4]. Multiple studies have shown that PTx is very effective in lowering PTH levels in ESRD patients, and successful PTx may reduce the risk of all-cause and cardiovascular mortality in ESRD patients with severe uncontrolled HPT [5,6]. Since secondary HPT may regress in up to 57% of patients with correction of mineral homeostasis after successful KTx, PTx is often postponed in patients listed for transplantation [7][8][9]. On the other hand, persistently elevated parathyroid hormone (PTH) levels cause resistance of the parathyroid to serum calcium levels due to reduced expression of the calcium-sensing receptor [10,11]. This results in tertiary HPT: autonomous production of PTH with subsequent hypercalcemia, which occurs in 25-50% of patients after KTx [12,13]. PTx is considered the only definitive treatment for tertiary HPT [14]. The introduction of calcimimetics in 2004 has reduced PTx rates and contributed to prolonged exposure to elevated serum PTH levels [15]. Previous studies suggest that a higher pre-KTx PTH level is associated with accelerated estimated glomerular filtration rate (eGFR) decline after KTx, and an increased risk of graft failure [3,16,17]. Furthermore, previous studies suggested a transient deterioration of renal graft function after PTx [16,18,19]. These previous studies report discordant results and arguments regarding the optimal timing of PTx in relation to KTx. In this study, we aimed to assess the impact of the sequence of KTx and PTx on the course of post-KTx renal function in a large multicenter cohort of ESRD-related HPT patients with long-term follow-up.

Study design
We performed a multicenter retrospective cohort study in patients who underwent both KTx  We evaluated medical records of all patients who underwent both a KTx and a PTx in these centers between 1994 and 2015. All patients were C18 years and diagnosed with ESRD-related HPT. Patients were divided into two groups according to treatment sequence: the PTxKTx group, who underwent PTx before KTx, and the KTxPTx group, who underwent PTx after KTx. When patients received more than one kidney transplant, only the first KTx was taken into account.
This study was approved by the local medical ethical committee of all participating centers (METc 2014/077). The study was performed according to the Helsinki Ethical Principles.

Data collection
For all patients, we collected cause of ESRD, pre-transplant dialysis status (preemptive or dialysis), history of diabetes mellitus, donor age and sex, cold and warm ischemia times, number of HLA mismatches, primary nonfunction (PNF), type of PTx (subtotal PTx or total PTx with autotransplantation), and biochemistry. PNF was defined as an eGFR \10 ml/min/1.73 m 2 at 3 months after KTx. Patients who reached ESRD (dialysis or re-transplantation) during follow-up were denoted as having an eGFR of 0 ml/min/1.73 m 2 until end of follow-up at 5 years; patients who died during follow-up were censored. The following biochemical measurements were recorded: serum calcium, PTH, albumin, and creatinine prior to KTx and PTx and at 3 months, 6 months, 1 year, 3 years, and 5 years after both KTx and PTx. For patients who underwent KTx after 2014, only 3-year follow-up data are available. eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) in ml/ min/1.73 m 2 [20]. Serum calcium and albumin were measured using routine laboratory techniques. Serum calcium was corrected for albumin levels according to the following formula: adjusted total calcium (mg/dL) = measured calcium (mg/dL) ?

Primary and secondary endpoints
To analyze the impact of the timing of PTx on graft function, the primary endpoint was eGFR at 5 years after transplantation. We also analyzed serum-corrected calcium, and PTH, graft failure and post-PTx complications, including temporary palsy of the recurrent laryngeal nerve (RLN), surgical site problems (SSP, including hematoma and infection), hospital-acquired pneumonia (HAP), intensive care unit (ICU) admission, and temporary hypocalcaemia.

Data analyses and statistics
A power analysis was performed based on results of a previous comparable study [21]. With 80% power and a two-sided a = 0.05, a simple size of n = 120 is required to detect a 20% difference in eGFR post-KTx. Statistical analysis was performed using SPSS Statistics version 24.0 (IBM Corporation, Armonk, NY, USA); a P value of\0.05 was considered statistically significant.
Patient characteristics were compared between the two groups (PTxKTx vs. KTxPTx) using Mann-Whitney U test and Pearson's Chi-square test where appropriate. Continuous variables were reported as mean ± standard error of the mean (SEM) or median with interquartile range (IQR). Categorical variables were expressed as number (n) and percentage (%).
To further study the impact of treatment sequence on graft function over time, we established generalized estimating equations (GEE) models with an exchangeable correlations structure. Based on previous literature, recipient age and sex, donor age and sex, type of donor (living vs. deceased), total number of HLA mismatches, type of PTx (total vs. subtotal), cold ischemia time, and pretransplant dialysis status (preemptive or dialysis) were defined as potential important confounders prior to analysis and were adjusted for in the multivariable GEE model. Furthermore, baseline variables with P values \0.2 in the univariate GEE analyses were included in multivariable GEE model. Results of the GEE model are displayed as estimates of the effects (B) and 95% confidence interval (CI) with P value. In a sub-analysis, we also evaluated the impact on eGFR in patients who underwent PTx shortly after KTx (\1 year) or longer (C1 year) after KTx.

Patient characteristics
A total of 185 patients were included: 102 (55.1%) patients underwent PTx before KTx (PTxKTx group), while 83 (44.9%) underwent PTx after KTx (KTxPTx group) (Fig. 1). Baseline patient and transplant characteristics are presented in Table 1. Patients within the KTxPTx group were significantly younger at the time of KTx than patients in the PTxKTx group. Eight patients (4.3%) underwent preemptive KTx; these patients were well equally distributed among the groups (n = 3 vs. n = 5, p = 0.31   pmol/L, p \0.001).
Calcimimetics were used at some point during follow-up in 31.4% of the patients in the PTxKTx group, compared to 20.5% of the patients in de KTxPTx group (p = 0.18). Ten patients (10.1%) developed primary non-function after KTx in the PTxKTx group, compared to 17.9% in the KTxPTx group (p = 0.13), and were excluded from further analysis.

Impact of PTx timing on post-KTx eGFR
The unadjusted GEE model showed that the timing of PTx was not associated with graft function over time (mean difference -1.0 ml/min/1.73 m 2 , 95% confidence interval [CI] -8.4 to 6.4, p = 0.79), Table 2). In a model adjusted for donor variables including donor type, donor gender, total number of HLA mismatches, whether the transplantation was preemptive or post-dialysis, donor age, and cold ischemia time, the mean difference in eGFR was -4.8 ml/ min/1.73 m 2 (96% CI -15.4 to 5.7, p = 0.37). Finally, we constructed a third model adjusted for pre-defined potentially relevant covariates and all baseline variables with a P value of \0.2 in univariate analysis. This analysis also showed that the course of eGFR over time was not significantly different between patients who underwent PTx before KTx or after KTx (Table 2).

Discussion
This large multicenter cohort study showed that in patients with ESRD-related HPT, the timing of parathyroidectomy (PTx) before or after kidney transplantation (KTx) does not independently impact the long-term course of kidney function after KTx.
This finding is in line with a previous single-center study that reported no significant difference in graft function at various time points after KTx between patients who underwent PTx before or after KTx [21]. With 185 patients, our study population is almost twice as large and meets the pre-specified sample size sufficient to detect a 20% difference in eGFR after PTx. Moreover, our followup was 5 years compared to 12 months in the previous study. Our results are at variance with two smaller retrospective studies. One study (Schwarz et al. [19]) investigating 76 kidney transplant recipients who underwent PTx showed that nearly half of these patients had an eGFR decline of more than 20%; however, renal function returned to pre-PTx values at 1 year post-PTx, in line with our results. In this study, patients who had this deterioration of graft function had a greater delta PTH decline after PTx compared to those without a deterioration [19]. In the current study, no significant differences were found between the stable eGFR group and the patients who had an eGFR decline C25%. This is likely due to the small number of patients with such eGFR decline post-PTx in our cohort. Parikh et al. also reported a significant but transient decline in eGFR post-PTx [16]. Another recent study by Littbarski et al. [22] suggested that, particularly early (\1 year) after transplantation, PTx contributes to renal function loss. Despite the larger sample size in our study We found that pre-PTx PTH levels were significantly lower in the group of patients who underwent PTx after KTx. This could be partly explained by improved renal function resulting in at least partly restored mineral homeostasis [23]. Alternatively, the difference in pre-KTx PTH values could result from the fact that patients with more severe HPT were more likely to undergo PTx prior to receiving a kidney transplant.
Factors determining the post-transplant eGFR course include donor and recipient age and sex, history of diabetes, cold ischemia time, total number of HLA mismatches, and donor type [24]. Our results indicate that the timing of PTx in relation to KTx does not importantly impact the course of renal function. Therefore, other factors than impact on graft function should be taken into account when deciding on a treatment plan and sequence in   [25]. HPT is associated with an increased risk of cardiovascular disease events and mortality [3,26]. Therefore, when KTx is expected in the near future and the abovementioned factors are taken into account, we suggest to be reticent to proceed with PTx. Based on our results, it seems safe for patients with persistent HPT after KTx to undergo PTx, at least regarding renal function and comparable complication rates between the two groups (Supplementary Table 2). Additionally, there was no significant difference in HPT recurrence rate between both study groups. In a recent study, male sex and white race were predictors of a more pronounced decline in PTH 2 years after KTx, whereas obesity, dialysis vintage, and delayed graft failure were risk factors for persisting HPT after KTx [7]. Studies aimed at identifying predictors of PTH normalization after KTx support a personalized approach toward the timing of PTx in relation to KTx.
Several limitations of our study should be addressed. Inherent to the retrospective nature of our study, selection bias and residual confounding cannot be excluded. Many patient-specific factors have likely driven the decision on PTx timing. We were therefore not able to determine the exact indication criteria for PTx in this multicenter retrospective study cohort. It is, however, probable that patients with more severe disease were more likely to have undergone parathyroidectomy prior to their kidney transplantation compared to the patients with less severe disease. Prospective studies are needed to provide more solid evidence. Second, our results obtained in four centers in the Netherlands might not be extrapolated to other countries with different healthcare infrastructure and guidelines. Third, we could not take the introduction of calcimimetics into account, which has significantly changed management of HPT [15].
In conclusion, we found that the sequence of KTx and PTx does not independently influence post-transplant graft function. We suggest that PTx can be safely performed after KTx in patients with persistent HPT. Therefore, we support the approach to postpone PTx until after KTx if transplantation is expected within a reasonable timeframe in order to allow spontaneous HPT regression, which occurs in a considerable proportion of patients. Our findings support a personalized approach for HPT patients listed for kidney transplantation.